A Study Manual for: The PACE Study · from 2014 to 2015 who were either newly, returning, or...

1

A Study Manual for:

The PACE Study

Post Affordable Care Act: Evaluation of Community Health Centers

AHRQ R01HS024270

NCT 02657499

January 1. 2019

A partnership between: Oregon Health & Science University

OCHIN, Inc.

2

Manual of Procedures Summary ................................................................................................. 3 Overview of Study ...................................................................................................................... 4

Funding ................................................................................................................................... 4 Table 1. Study Staff. ............................................................................................................ 4

ClinicalTrials.gov Protocol Submission ................................................................................... 4 Specific Aims.......................................................................................................................... 4 Data Sources ........................................................................................................................... 5

OCHIN EHR Data. ........................................................................................................... 5 Linkage between EHR and Medicaid administrative data. ................................................ 5

Study Population and Study Period ......................................................................................... 5 Table 2. Study population stratified by expansion status and state ....................................... 6

Stakeholder Engagement ......................................................................................................... 6 Data Analyses ............................................................................................................................. 7 Project Bibliography ................................................................................................................. 21 Regulatory Requirements .......................................................................................................... 25

Institutional Review Board (IRB) .......................................................................................... 25 Approval ........................................................................................................................... 25 Protocol ............................................................................................................................. 25

Data Use Agreements and Amendments ................................................................................ 25 Table 3. Data Use Agreements .......................................................................................... 25

Decision Log ............................................................................................................................ 26 Table 4. Decision Log ....................................................................................................... 26

Appendix 1. IRB Approval ....................................................................................................... 29 Appendix 2. Project Protocol .................................................................................................... 31 Appendix 3. Data Use Agreements ............................................................................................ 37

3

Manual of Procedures Summary

This manual of procedures provides an overview of the PACE project, and is meant to serve as a resource to facilitate the replication of this study in other settings. An overview of the study describes the project aims, as well as information about the study team, funding, and data sources. Further sections describe the regulatory aspects of the study, major project decisions, and the analyses conducted.

4

Overview of Study

Funding This work was supported by the Agency for Healthcare Research and Quality grant number R01HS024270, and was funded September 2015 through September 2017.

Table 1. Study Staff. Name Project Role Institution Jennifer DeVoe, MD, DPhil Principal Investigator OHSU Heather Angier, PhD, MPH Co-Project Director, Co-Investigator OHSU Nathalie Huguet, PhD Co-Project Director, Co-Investigator OHSU Miguel Marino, PhD Co-Investigator, Biostatistician OHSU Lewis Raynor, PhD, MS, MPH Co-Investigator OCHIN Megan Hoopes, MPH Data Analyst - Site PI (Year 2- NCE) OCHIN Heather Holderness, MPH Project Manager OHSU Nate Warren, MPH Project Manager OCHIN Jean O’Malley, MPH Data Analyst OHSU Rachel Springer, MS Data Analyst OHSU Steele Valenzuela, MS Data Analyst OHSU David Ezekiel-Herrera, MS Data Analyst OHSU Teresa Schmidt, MS Data Analyst OCHIN Pedro Rivera-Ortega Data Programmer Year 1 OCHIN Allison O’Neil Data Analyst Year 1 OCHIN Carlyn Hood Project Manager Year 1 OCHIN Kaleb Keaton Project Manager Year 1 OHSU

ClinicalTrials.gov Protocol Submission ClinicalTrials.gov Identifier: NCT02657499 Specific Aims Aim 1. Compare pre-post health insurance status, primary care, mental health, and dental visits, and receipt of preventive services, as well as changes in payer mix among ADVANCE CHCs in states that did and did not expand Medicaid. Hypothesis 1a: CHCs in expansion states will experience an increase in overall visits and visits paid by Medicaid, relative to non-expansion states. Hypothesis 1b: The percentage of Medicaid visits in CHCs' overall payer mix will increase more significantly in expansion-states relative to non-expansion states. Aim 2. Examine pre-post utilization of CHC services (including receipt of preventive services) by newly insured compared to already insured and uninsured patients. Hypothesis 2: Newly insured patients will have a significant increase in overall CHC visits compared to (a) already insured and (b) uninsured patients. Aim 3: Measure pre-post changes in overall utilization of healthcare services and Oregon Medicaid expenditures among newly insured compared to already insured patients.

5

Hypothesis 3a: There will be an increase in utilization of CHC services and a decrease in use of services external to the CHCs (e.g., emergency department) for newly insured patients, but these utilization rates will remain the same for already insured patients. Hypothesis 3b: There will be an initial increase, then plateau, in overall Medicaid expenditures among newly insured relative to already insured patients. Data Sources

ADVANCE EHR Data. We used electronic health record (EHR) data from the Accelerating Data Value Across a National Community Health Center Network (ADVANCE) clinical data research network (CDRN) data warehouse. The ADVANCE CDRN is a unique ‘community laboratory’ for research with underrepresented populations that includes patients receiving care in safety net clinics. The ADVANCE CDRN data warehouse includes integrated longitudinal EHR data from several organizations. Data from two ADVANCE partners, OCHIN and Health Choice Network (HCN), were used in the PACE study. From the ADVANCE CDRN, we included CHC members ‘live’ on their EHR as of January 1, 2013: 225 primary care CHCs in 10 states that expanded Medicaid as of 1/1/2014 (California, Hawaii, Maryland, Minnesota, New Mexico, Ohio, Oregon, Rhode Island, Washington, and Wisconsin) and 134 primary care CHCs in 6 non-expansion states (Florida, Kansas, Missouri, North Carolina, Texas, Montana). We included Wisconsin as an expansion state because although they did not expand Medicaid to 138% FPL they opened enrollment to adults with eligibility criteria of 100% FPL and therefore behaved more like an expansion state. Montana did not expand Medicaid until after our study period (expanded 1/1/2016), which included a one-year pre- (1/1/13-12/31/13) and a two-year post-period (1/1/2014-12/31/2015). Analyses included >4 million ambulatory visits among patients aged 19-64 (expansion states: 2.6 million visits from 499,719 patients; non-expansion states: 1.5 million visits from 370,600 patients).

Linkage between EHR and Medicaid administrative data. Oregon Medicaid administrative data was linked to OCHIN Oregon clinics in the ADVANCE CDRN to assess the receipt of healthcare services outside the Oregon CHCs and all healthcare expenditures. Study Population and Study Period The overall study population was adults ages 19-64 throughout the study period, 2012-2015. Included patients had ≥1 ambulatory visit at an eligible facility in the study period. Included facilities provided adult primary care and were ‘live’ on OCHIN’s or HCN’s EHR system by the start of the study period (1/1/2012). EHR data were extracted on all eligible patients, encounters, and facilities.

Overall, study analyses included >4 million ambulatory visits among patients aged 19-64 (expansion states: 2.6 million visits from 499,719 patients; non-expansion states: 1.5 million visits from 370,600 patients).

6

We examined differences in utilization among patients newly gaining Medicaid coverage (newly insured), as compared to utilization among individuals who are already insured by Medicaid (already insured) and those who remain uninsured (uninsured). Additionally, in subpopulation analyses restricted to the state of Oregon (an expansion state), we link EHR data from Oregon CHC clinics to Oregon’s Medicaid administrative claims data to investigate post-expansion overall healthcare utilization and Medicaid expenditures among the newly insured as compared to the already insured. Inclusion and exclusion criteria were modified for some analyses; see below for details on specific paper analyses.

Table 2. Study population stratified by expansion status and state 2014 Expansion status State N facilities N patients Expansion California 48 163,686 Expansion Hawaii 2 10,075 Expansion Maryland 24 15,297 Expansion Minnesota 4 12,725 Expansion New Mexico 25 63,488 Expansion Nevada 3 2,326 Expansion Ohio 15 55,537 Expansion Oregon 121 261,166 Expansion Rhode Island 59 38,946 Expansion Washington 6 24,585 Non-expansion Alaska* 1 3,820 Non-expansion Florida 169 412,138 Non-expansion Indiana** 10 16,200 Non-expansion Kansas 9 8,047 Non-expansion Missouri 3 16,461 Non-expansion Montana 6 13,155 Non-expansion North Carolina 21 39,726 Non-expansion Texas 5 11,383 Non-expansion Wisconsin*** 8 20,593 *Alaska expanded Medicaid to 138% FPL on 9/1/2015 and was excluded from some analyses; **Indiana expanded Medicaid to 138% FPL on 2/1/2015 and was excluded from some analyses; ***Wisconsin expanded Medicaid from 0% to 100% on 1/1/2014 and therefore more closely resembled an ACA expansion state. Wisconsin was considered an expansion state in analyses.

Stakeholder Engagement The project team engaged a Patient Investigator (Kaye Dickerson) and a Clinician Investigator (Dr. Andrew Suchocki, MD) throughout the duration of the project. These stakeholders advised the project team during the data analysis, manuscript development, and dissemination phases of the project. Their involvement was facilitated through ongoing training about the research process, collaboration with research team about data analyses challenges, data interpretation,

7

and manuscript development, and dissemination of research aims through presentations and blogging.

Data Analyses

Medicaid Expansion Produces Long-Term Impact on Insurance Coverage Rates in Community Health Centers

Full Citation Huguet, N., Hoopes, M. J., Angier, H., Marino, M., Holderness, H., & DeVoe, J. E. (2017) Medicaid Expansion Produces Long-Term Impact on Insurance Coverage Rates in Community Health Centers. J Prim Care Community Health. 8(4):206-212.

Objective Assess change in insurance coverage rates pre- and post-ACA in Medicaid expansion and non-expansion states for community health center patients.

Study Population N=875,571 patients; Inclusion: ages 19-64 with ≥ 1 visit between 2012 and 2015. Setting: 412 primary care CHCs in 9 expansion and 4 non-expansion states.

Exclusion Criteria Pregnant women Independent Variables

Whether state had taken Medicaid expansion as of January 1, 2014.

Dependent Variables

Health care delivery: rates of billed encounters (all, primary care, new patient, established patient), receipt of preventive services (preventive medicine visits, immunizations, medications ordered). 1. Insurance type: uninsured, Medicaid-insured, privately-

insured. Covariates Clinic-level: distributions of sex, age, race/ethnicity, primary

language, federal poverty level. State-level: marketplace type, 2013 minimum wage, 2013 uninsured rate, 2013 unemployment rate.

Analysis Plan Post- versus pre-expansion rate ratios within expansion group; difference-in-difference ratios (comparing post- versus pre-period changes between expansion states vs non-expansion states); and second year post- (2015) versus first year post-ACA (2014) rate ratios within expansion group.

Data Set Name OCHIN: e:\sasroot\PPDA\P_Huguet_24mo Code Location OCHIN SAS server

8

Oregon Medicaid Expenditures After the 2014 Affordable Care Act Medicaid Expansion: Over-time Differences Among New, Returning, and Continuously Insured Enrollees

Full Citation Springer, R., Marino, M., O'Malley, J. P., Lindner, S., Huguet, N., & DeVoe, J. E. (2018). Oregon Medicaid Expenditures After the 2014 Affordable Care Act Medicaid Expansion: Over-time Differences Among New, Returning, and Continuously Insured Enrollees. Med Care, 56(5), 394-402.

Objective To assess health care expenditures among Medicaid enrollees in the 24 months after Oregon’s 2014 Medicaid expansions and examine whether expenditure patterns were different among the newly, returning, and continuously insured.

Study Population Oregon adult Medicaid beneficiaries insured continuously from 2014 to 2015 who were either newly, returning, or continuously insured.

Exclusion Criteria Our study population included adults aged 19-64 insured by Oregon Medicaid continuously from January 2014 through December 2015. We excluded patients without coverage on 1/1/2014 and those with coverage gaps during the study period. We excluded patients with dual Medicaid and Medicare eligibility and whose 2014-2015 eligibility did not depend on FPL (e.g. pregnant women). Of 622,513 adults aged 19-64 with any 2014 Medicaid enrollment, 230,602 (37%) remained in our sample.

Independent Variables

Insurance group, time, and their interaction terms, mean monthly expenditures.

Dependent Variables

standardized expenditures for inpatient care, prescription drugs, and total outpatient care in addition to subdivisions of outpatient: outpatient claims were classified as either emergency department, dental, mental and behavioral health, primary care evaluation, management, and procedures, specialist evaluation, management, and procedures, and primary care/specialist imaging or tests based on a hierarchical system involving procedure code, provider type, provider specialty, place of service, and associated costs.

Covariates The patient’s age, sex, racial and ethnic background, rural setting, and comorbidity level as assessed by the enhanced Charlson comorbidity index score.

Analysis Plan Retrospective cohort study using inverse-propensity weights to adjust for differences between groups. We summarized enrollee

9

utilization outcomes as mean monthly expenditures, running a separate (two-part) hurdle model for each type.

Data Set Name NPAT, INPATIENT_Q42015, OUTPAT, OUTPATIENT_2015Q4, RX , RX_Q42015, SAMPLE, SAMPLE_with_weights FY_2015_FR_AND_CN_Table 5, Nov 2013 Fee Schedule – Excel, January 2014 Behavioral Health Fee Schedule, January 2014 Mental Health Fee Schedule – Excel

Code Location Box at OHSU Medicaid Coverage Accuracy in Electronic Health Records

Full Citation Marino, M., H. Angier, S. Valenzuela, M. Hoopes, M. Killerby, B. E. Blackburn, N. Huguet, J. Heintzman, B. Hatch, J. P. O' Malley and J. E. DeVoe (2018). Medicaid Coverage Accuracy in Electronic Health Records. Prev Med Reports 11: 297–304.

Objective To evaluate the validity of EHR data for monitoring longitudinal Medicaid coverage and assess variation by patient demographics, visit types, and clinic characteristics.

Study Population Oregon Medicaid patients aged 18-64 years with a Medicaid ID or Medicaid insurance recorded in the OCHIN EHR ≥1 billed health care visit (excluding dental) from 184 Oregon CHCs in the OCHIN network linked to state Medicaid records during the study time period (1/1/2013–12/31/2014). N=135,514.

Exclusion Criteria Patients with no Medicaid record on EHR who also did not have a Medicaid ID.


Insurance cohort: (1) continuously Medicaid (Medicaid recorded at all visits); (2) continuously not Medicaid (Medicaid not recorded at any visit (patients could have Medicare, private, VA/Military, worker’s compensation, or no coverage); (3) Gained Medicaid: Medicaid not recorded for all visits in 2013 and Medicaid recorded at all visits in 2014; and (4) Discontinuously Medicaid: Any combination of visit coverage that did not follow the definitions above.

Dependent Variables

The primary outcome was agreement between EHR and Medicaid data in assigning patients to one of the four cohorts.

Covariates Patient demographics (sex, age, race, ethnicity, language, household income represented as % FPL, urban/rural, number of common chronic conditions, and number of encounters), visit and provider types, and clinic-specific (department type and customers of OCHIN's billing service) characteristics. We included the following common chronic conditions assessed

10

from diagnostic codes in the EHR: hypertension, diabetes, coronary artery disease, lipid disorder, and asthma/chronic obstructive pulmonary disorder.

Analysis Plan Visit-level: agreement, prevalence-adjusted bias-adjusted kappa (PABAK), sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). Patient-level: 4x4 cross-tabulation, of insurance cohorts; two-stage logistic regression.

Data Set Name PPDA Projects à InsuranceValidation à Data à CLEAN à patients.csv, patient_pre.csv, patient_post.csv, patient_ch.csv, encs_pre.csv, encs_post.csv, agree_no_final.csv Code is included under: PPDA Projects à InsuranceValidation à Final Code

Code Location Box at OHSU

In Low-Income Latino Patients, Post-Affordable Care Act Insurance Disparities May Be Reduced Even More than Broader National Estimates: Evidence from Oregon

Full Citation Heintzman, J., S. R. Bailey, J. DeVoe, S. Cowburn, T. Kapka, T. V. Duong and M. Marino (2017). In Low-Income Latino Patients, Post-Affordable Care Act Insurance Disparities May Be Reduced Even More than Broader National Estimates: Evidence from Oregon. J Racial Ethn Health Disparities 4(3): 329-336.

Objective To evaluate the association of Hispanic/Latino ethnicity and Spanish language preference with insurance status among a cohort of patients accessing community health centers (CHCs) in Oregon; prior to ACA implementation (2013) and 1-year post-ACA (2014). To determine if disparities in insurance coverage existed by race/ethnicity and language preference prior to ACA. If so, were such disparities mitigated by ACA implementation.

Study Population N=49,392; Adults aged 21-79 in Oregon community health centers (CHCs) who had low income (<100% FPL), and at least 1 primary care encounter at a study clinic in 2009-2013, had to have at least 1 face to face primary care encounter.

Exclusion Criteria Independent Variables

Race/ethnicity, preferred language

Dependent Variables

Insurance status: Medicaid, Medicare, private insurance, uninsured.

11

Categorized patients into the Medicaid and Medicare categories if patient had one or more encounters with these insurances.

Covariates Age at beginning of study, sex, total number of primary care office visits in study period.

Analysis Plan Chi-square for patient demographics; ANOVA to compare patient demographics by insurance group; GEE to account for correlation of patients nested with CHCs; DID/GEE to assess changes in odds of uninsured from pre-to post-ACA between Hispanic/Latino/Spanish, Hispanic/Latino/English, and non-Hispanic White cohorts.

Data Set Name PPDA Projects à PACE à HeintzmanACA à 2014data.xls, heintzman_k_post_aca_comparison.xls Code is included under: PPDA Projects à PACE à Heintzman ACA à Code

Code Location Box at OHSU Medicaid's Impact on Chronic Disease Biomarkers: A Cohort Study of Community Health Center Patients.

Full Citation Hatch, B., M. Marino, M. Killerby, H. Angier, S. R. Bailey, J. Heintzman, J. P. O' Malley and J. E. DeVoe (2017). Medicaid's Impact on Chronic Disease Biomarkers: A Cohort Study of Community Health Center Patients. J Gen Intern Med 32(8): 940-947.

Objective To assess changes in biomarkers of chronic disease among community health center (CHC) patients who gained Medicaid coverage with the Oregon Medicaid expansion (2008–2011).

Study Population OCHIN Medicaid patients within 3 chronic disease cohorts: patients with diabetes (N=608); patients with hypertension patients (N=1244); patients with hyperlipidemia patients (N=546). Patients gained Medicaid coverage between 2008-2011: Patients aged 1-+64 years; Non-pregnant, non-deceased; Uncontrolled biomarker result in EHR within 6 months before/after gained Medicaid coverage.


Insurance status, time, “spline” (slope change at 6 months) interactions between time and insurance and spline and insurance.

Dependent Variables

Biomarkers: 1. HbA1c

12

2. Systolic blood pressure 3. Diastolic blood pressure 4. Lipids

Covariates federal poverty level, comorbidities Analysis Plan Time to event assessing time from uncontrolled baseline

measurement to follow up controlled measurement; Longitudinal analysis modeling the mean biomarker value over time; Logistic regression to estimate the odds of having a disease-specific medication ordered during study.

Data Set Name PPDA Projects à PACE à CARDIACBiomarker à Data à all the files in there Code is under: PPDA Projects à PACE à CARDIACBiomarker à Rcode


13

Uninsured Primary Care Visit Disparities Under the Affordable Care Act Full Citation Angier, H., M. Hoopes, M. Marino, N. Huguet, E. A. Jacobs, J.

Heintzman, H. Holderness, C. M. Hood and J. E. DeVoe (2017). Uninsured Primary Care Visit Disparities Under the Affordable Care Act. Ann Fam Med 15(5): 434-442.

Objective To assess changes in insurance coverage at community health center (CHC) visits after the Patient Protection and Affordable Care Act (ACA) Medicaid expansion by race and ethnicity

Study Population 1. N=870,319 2. 359 CHCs 3. CHC patients aged 19-64 years 4. Expansion states as of 1-1-14: California, 5. Hawaii, Maryland, Minnesota, New Mexico, Ohio,

Oregon, Rhode Island, Washington, and Wisconsin 6. Non-expansion states as of 1-1-14: Florida, Kansas,

Missouri, North Carolina, Texas, Montana Exclusion Criteria Pregnant women Independent Variables

Expansion status (Medicaid expansion vs. non-expansion)

Dependent Variables

Health insurance type at each visit (Medicaid-insured, uninsured, privately-insured)

Covariates Analysis Plan 1. Difference-in-difference (DID)

2. Difference-in-difference-in-difference (DDD) 3. GEE models

Data Set Name HA_disp_analysis_long.sas7bdat HA_disp_monthly_analysis_set.sas7bdat

Code Location OCHIN archive server

14

A Cohort Study of Public Health Insurance Coverage Loss among Oregon Adolescents Full Citation Angier, H., C. J. Tillotson, L. S. Wallace, M. Marino, J. P. O' Malley,

A. Sumic, L. Baker, C. Nelson, N. Huguet, A. Suchocki, H. Holderness and J. E. DeVoe (2018). A Cohort Study of Public Health Insurance Coverage Loss among Oregon Adolescents. J Health Dispar Res 11(1).

Objective The purpose of this study was to identify the sociodemographics of adolescents transitioning to young adulthood who lost public health insurance which could inform future health policies for this vulnerable population.

Study Population Oregon adolescents (17-19 years of age) with public coverage [January 1, 2011-December 31, 2013 (n=51,988).

Exclusion Criteria Missing information on urban/rural residency status (n=4,025) and/or family income (n=973).


Sociodemographic characteristics: age, sex, race, ethnicity, language, percentage of federal poverty level based on family income and size, history of a healthcare visit to a community health center included in the OCHIN network during the study period (2011-2013), and urbanicity.

Dependent Variables

OHP Coverage

Covariates Analysis Plan Time to Event: to determine association of coverage loss with

sociodemographic characteristics. Data Set Name analysis_nogaps_01202016.sas7bdat

dmapochin13_nogaps_01202016.sas7bdat Code is under: R:\Active projects\IMPACCT\Reports\Data analysis and results\Teen Transition\Carrie\SAS Code and Programs\5d_MV Models nogaps 01112016.sas

Code Location OCHIN archive server

A New Role for Primary Care Teams in the United States After “Obamacare”: Track and Improve Health Insurance Coverage Rates.

Full Citation DeVoe, J. E., H. Angier, M. Hoopes and R. Gold (2016). A New Role for Primary Care Teams in the United States After “Obamacare”: Track and Improve Health Insurance Coverage Rates. Family Medicine and Community Health 4(4): 63-67(65).

Objective Describe efforts to longitudinally track health insurance rates using data from the electronic health record (EHR) of a primary care network and to use these data to support practice-based insurance outreach and assistance.

15

Study Population OCHIN member clinics and patients. Exclusion Criteria Not an OCHIN member clinic or patient. Independent Variables

NA

Dependent Variables

Insurance coverage in EHR.

Covariates NA Analysis Plan NA Data Set Name NA Code Location NA

The Impact of the Affordable Care Act Medicaid expansion on visit rates for a patient population with diabetes or pre-diabetes in safety net health centers

Full Citation Huguet, N., R. Springer, M. Marino, H. Angier, M. Hoopes, H. Holderness and J. E. DeVoe. (2018) The Impact of the Affordable Care Act Medicaid expansion on visit rates for a patient population with diabetes or pre-diabetes in safety net health centers. J Am Board Fam Med. 31(6):905-916

Objective To compare clinic-level uninsured, Medicaid-insured, and privately insured visit rates within and between expansion and non-expansion states prior to and after the ACA Medicaid expansion among the three cohorts of patient populations; and, to assess whether there was a change in clinic-level overall, primary care, preventive care visits, and diabetes screening rates in expansion versus non-expansion states from pre- to post-ACA Medicaid expansion.

Study Population 198 primary care community health centers: Non-pregnant patients aged 19-64 with ≥1 ambulatory visit between 01/01/2012-12/31/2015 (n=483,912 in expansion states; n=388,466 in non-expansion states).


Expansion status

Dependent Variables

1. Insurance coverage 2. Healthcare services

Covariates Health insurance status, sociodemographic variables, urban/rural, state-level factors.

Analysis Plan • Rate Ratios • Difference in Difference • GEE/Poison models with sandwich variance estimators • Clustered models by CHC

16

• Exchangeable covariance structure • Sensitivity analysis

Data Set Name STATES, ELIGIBLE FACILITIES, ENCOUNTERS_AV1215_pipe, PAT_CONSORT_08312016, FAC_URBRUR

Code Location BOX at OHSU

Prescription Opioid Use Patterns, Use Disorder Diagnoses, and Addiction Treatment Receipt after the 2014 Medicaid Expansion

Full Citation Springer, R., M. Marino, S. Bailey, H. Angier, J. O’Malley, M. Hoopes, S. Lindner, J. DeVoe and N. Huguet. Prescription Opioid Use Patterns, Use Disorder Diagnoses, and Addiction Treatment Receipt after the 2014 Medicaid Expansion. Resubmitted 12/18 to Addiction.

Objective To examine rates of opioid prescribing, Opioid Use Disorder (OUD) prevalence, and OUD treatment among Oregon Medicaid enrollees after the ACA Medicaid expansion; and to understand how insurance history and dose type are related to OUD diagnosis rates, as well as whether patients with OUD in the expansion population (of new enrollees and those ‘returning’ after coverage gaps) had different MAT participation rates than those previously continuously enrolled.

Study Population Adults aged 19+64; continuously insured by Oregon Medicaid 1/2014-12/31/2015. N=225,295.

Exclusion Criteria Patients with coverage gaps during study period; dual eligibles; hospice care; cancer diagnosis other than non-melanoma skin cancer.


1. Insurance group 2. Episode type

Dependent Variables

1. Opioid filled prescription 2. Documented diagnosis of OUD 3. Chronic opioid use 4. Receipt of medication assisted treatment

Covariates Age, sex, race, urban/rural, co-morbidities, zip code, federal poverty level, employment, and number of episodes

Analysis Plan 1. Inverse propensity weighting 2. Absolute standardized mean difference 3. Prevalences 4. Logistic regression 5. Cox proportional hazard 6. Multinomial logistic regression

Data Set Name OUTPAT, OUTPATIENT_2015Q4, RX , RX_Q42015, SAMPLE_FLAGGED_FOR_EXPORT

17

hospice_providers, mat_codes, opioid_ndcs_2016

Code Location Box at OHSU Following uninsured patients through Medicaid Expansion: healthcare utilization and diagnosed conditions

Full Citation Huguet, N., S. Valenzuela, M. Marino, H. Angier, B. Hatch, M. Hoopes and J. E. DeVoe. Following uninsured patients through Medicaid Expansion: healthcare utilization and diagnosed conditions. Submitted Ann Fam Med 11/18.

Objective Assess healthcare utilization and diagnosed health conditions among a cohort of community health center patients who were uninsured pre-ACA and follow them 24 months post-Affordable Care Act (ACA).

Study Population All non-pregnant patients without insurance pre-ACA, aged 19-64 during the study period, with ≥1 ambulatory visit pre-ACA from 300 primary care CHCs ‘live’ on their EHR system as of 1/1/2012 in 11 states. N=138,246.

Exclusion Criteria Pregnant women, insurance coverage prior to ACA. Independent Variables

1. Insurance group

Dependent Variables

1. Healthcare utilization 2. Health conditions

Covariates Age, sex, language, race/ethnicity, federal poverty level, and clinic location

Analysis Plan 1. Adjusted predicted probability 2. Mixed effects ordinal logistic regression 3. Random effects to account for clustering visits

Data Set Name PAT_CONSORT_08312016 encounters_AV1215_pipe FAC_URBRUR STATES PPDA_RUCA_CHARLSON PPDA_PREEXISTING_CONDITIONS Code is under: PPDA Projects -> Who_Remains_Uninsured -> draft4.Rmd


18

Utilization of Community Health Centers in Medicaid Expansion and Non-expansion States, 2013-2014.

Full Citation Hoopes, M. J., H. Angier, R. Gold, S. R. Bailey, N. Huguet, M. Marino and J. E. DeVoe (2016). Utilization of Community Health Centers in Medicaid Expansion and Non-expansion States, 2013-2014. J Ambul Care Manage 39(4): 290-298.

Objective Assess CHC utilization a full year before and after the implementation of Affordable Care Act (ACA) Medicaid expansions (24 months) and describes changes in visit type before and after implementation.

Study Population Adults 19-64 years, billed visits in pre expansion (1/1/13-12/31/13) and post expansion (1/1/14-12/31/14) OCHIN member clinics. N=401,988 patients.


Expansion status

Dependent Variables

1. Insurance status 2. Visit type

Covariates Sex, age, race/ethnicity, urban/rural, federal poverty level, minimum wage, unemployment, insurance exchange type.

Analysis Plan 1. GEE 2. Poisson models

Data Set Name OCHIN: e:\sasroot\CATCHUP\ACA 12 month Code Location OCHIN SAS server

Effect of Gaining Insurance Coverage on Smoking Cessation in Community Health Centers: A Cohort Study.

Full Citation Bailey, S. R., M. J. Hoopes, M. Marino, J. Heintzman, J. P. O'Malley, B. Hatch, H. Angier, S. P. Fortmann and J. E. DeVoe (2016). Effect of Gaining Insurance Coverage on Smoking Cessation in Community Health Centers: A Cohort Study. J Gen Intern Med 31(10): 1198-1205.

Objective To determine if uninsured community health center (CHC) patients who gained Medicaid coverage experienced greater primary care utilization, receive more cessation medication orders, and achieve higher quit rates, compared to continuously uninsured smokers.

Study Population CHC patients aged 19–64 years who gained Oregon Medicaid coverage between 2008 and 2011 after being uninsured for ≥ 6 months and who maintained this insurance for ≥ 6 months. Cohort of ‘current every day smoker’ or ‘current some day smoker.’

19

Exclusion Criteria Non-smoking, pregnant, non-living, private insurance, Medicare insurance.


Insurance Status

Dependent Variables

1. ‘Quit’ smoking status (yes/no) 2. Smoking cessation medication ordered (yes/no)

Covariates Usual source of care after in follow up period. Analysis Plan 1. Propensity score matching

2. Between group differences 3. GEE for clustering

Data Set Name Smoking_model_set_PSmatches.sas7bdat Code Location OCHIN archive server

Observational Study of Racial/Ethnic Disparities in Health Insurance and Differences in Visit Type for a Population of Patients with Diabetes after Medicaid Expansion.

Full Citation Angier, H., D. Ezekiel-Herrera, M. Marino, M. Hoopes, E. A. Jacobs, J. E. DeVoe and N. Huguet (In press). Observational Study of Racial/Ethnic Disparities in Health Insurance and Differences in Visit Type for a Population of Patients with Diabetes after Medicaid Expansion. J Health Care Poor Underserved.

Objective To understand the impact of the Affordable Care Act (ACA) Medicaid expansion on racial/ethnic health insurance disparities and differences in visit type for a population of patients with diabetes, by assessing changes pre- versus post-ACA in expansion compared with non-expansion states.

Study Population Adults aged 19-64, diabetes diagnosis any time during study period, CHCs in ADVANCE network with EHRs live as of January 2013.


Expansion status

Dependent Variables

1. Health insurance status 2. Healthcare visit types

a. Total b. Type

Covariates Sex, age, FPL, Charlson comorbidity index score, state-level factors

Analysis Plan 1. Incidence rates 2. GEE Poisson models 3. Difference in Difference

20

Data Set Name Analytic Data: PPDA Projects à PACE à Diabetes Disparities à Analytic Data Code: PPDA Projects à PACE à Diabetes Disparities à Code

Code Location Box at OHSU Assessing the prevalence of pre-existing conditions among community health center patients.

Full Citation Huguet, N., H. Angier, M. J. Hoopes, M. Marino, J. Heintzman, T. Schmidt and J. E. DeVoe. Assessing the prevalence of pre-existing conditions among community health center patients. Medical Care. Submitted 9/18

Objective To assess the prevalence of pre-existing conditions pre- versus post-Affordable Care Act (ACA) for community health center (CHC) patients who gained insurance

Study Population 386 CHCs in 19 states; 78,059 non-pregnant patients aged 19-64 uninsured at their last visit pre-ACA.


Medicaid expansion state status

Dependent Variables

Prevalence and types of pre-existing conditions by insurance status and race/ethnicity.

Covariates Analysis Plan 1. Prevalence

2. Within-group prevalence 3. GEE

Data Set Name OCHIN: e:\sasroot\PPDA\Huguet_preexist cond paper Code Location OCHIN SAS server

21

Project Bibliography

Manuscripts: 1. Huguet N, Hoopes MJ, Angier H, Marino M, Holderness H, DeVoe JE. Medicaid Expansion Produces

Long-Term Impact on Insurance Coverage Rates in Community Health Centers. J Prim Care Community Health. 2017;8(4):206-212. PMC5665709.

2. Angier H, Hoopes M, Marino M, Huguet N, Jacobs EA, Heintzman J, et al. Uninsured Primary Care

Visit Disparities Under the Affordable Care Act. Annals of Family Medicine. 2017;15(5):434-42. PMC5593726.

3. Hoopes MJ, Angier H, Gold R, Bailey SR, Huguet N, Marino M, et al. Utilization of Community Health

Centers in Medicaid Expansion and Nonexpansion States, 2013-2014. The Journal of Ambulatory Care Management. 2016;39(4):290-8. PMC4942402.

4. Springer R, Marino M, O'Malley JP, Lindner S, Huguet N, DeVoe JE. Oregon Medicaid Expenditures

After the 2014 Affordable Care Act Medicaid Expansion: Over-time Differences Among New, Returning, and Continuously Insured Enrollees. Med Care. 2018;56(5):394-402. PMC5893375.

5. Marino M, Angier H, Valenzuela S, Hoopes M, Killerby M, Blackburn BE, et al. Medicaid Coverage

Accuracy in Electronic Health Records. Prev Med Reports. 2018;11:297–304. PMC6082971. 6. Hatch B, Marino M, Killerby M, Angier H, Bailey SR, Heintzman J, et al. Medicaid's Impact on Chronic

Disease Biomarkers: A Cohort Study of Community Health Center Patients. Journal of General Internal Medicine. 2017;32(8):940-7. PMC5515790.

7. Huguet N, Springer R, Marino M, Angier H, Hoopes M, Holderness H, et al. The impact of the

Affordable Care Act Medicaid expansion on visit rates for diabetes in safety net health centers. Journal of the American Board of Family Medicine. 2018;31(6):905-16. PMID30413546

8. Angier H, Ezekiel-Herrera D, Marino M, Hoopes M, Jacobs EA, DeVoe JE, et al. Observational Study of Racial/Ethnic Disparities in Health Insurance and Differences in Visit Type for a Population of Patients with Diabetes after Medicaid Expansion. J Health Care Poor Underserved. In Press.

9. Angier H, Tillotson CJ, Wallace LS, Marino M, O' Malley JP, Sumic A, et al. A Cohort Study of Public Health Insurance Coverage Loss among Oregon Adolescents. J Health Dispar Res. 2018;11(1): 74-84.

10. Bailey SR, Hoopes MJ, Marino M, Heintzman J, O'Malley JP, Hatch B, et al. Effect of Gaining Insurance Coverage on Smoking Cessation in Community Health Centers: A Cohort Study. J Gen Intern Med. 2016;31(10):1198-205. PMC5023615.

11. DeVoe JE, Angier H, Hoopes M, Gold R. A New Role for Primary Care Teams in the United States After “Obamacare:” Track and Improve Health Insurance Coverage Rates. Family Medicine and Community Health. 2016;4(4):63-7(5). PMC5617364.

12. Heintzman J, Bailey SR, DeVoe J, Cowburn S, Kapka T, Duong TV, et al. In Low-Income Latino Patients, Post-Affordable Care Act Insurance Disparities May Be Reduced Even More than Broader

22

National Estimates: Evidence from Oregon. J Racial Ethn Health Disparities. 2017;4(3):329-36. PMC5075278.

Submitted Manuscripts: Springer R, Marino M, Bailey S, et al. Prescription Opioid Use Patterns, Use Disorder Diagnoses, and Addiction Treatment Receipt after the 2014 Medicaid Expansion. Addiction. Revised and Resubmitted 11/2018. Huguet N, Valenzuela S, Marino M, et al. Following uninsured patients through Medicaid Expansion: healthcare utilization and diagnosed conditions. Annals of Family Medicine. Revised and Resubmitted 11/2018. Huguet N, Angier H, Hoopes MJ, et al. Assessing the prevalence of pre-existing conditions among community health center patients. Medical Care. Submitted 9/2018. Bailey SR, Marino M, Ezekiel-Herrera D, et al. Tobacco Cessation in Affordable Care Act Medicaid Expansion States versus Non-Expansion States. Nicotine & Tobacco Research. Submitted 12/2018. Presentations: Valenzuela S, Huguet N, Springer R, Ezekiel-Herrera D, O' Malley JP, Marino M. Navigating Large-Scale Forest Plots Using R and Shiny. Poster Presentation presented at American Statistical Society Conference on Statistical Practice; Feb 15-17, 2018, 2017; Portland, OR. Springer R, Raynor L, Marino M, et al. Effects of Medicaid Expansion on Oregon Medicaid Patient Expenditures from 2014-2015. Oral Presentation on Completed Research presented at 45th annual meeting of the North American Primary Care Research Group; Nov 17-21, 2017; Montreal, Canada. Springer R, Marino M, Bailey SR, et al. Opioid Medication Expenditures in Oregon Medicaid Enrollees after the ACA Expansion. Poster presentation presented at AcademyHealth Annual Research Meeting; June 24-26, 2018; Seattle, WA. Springer R, Marino M, Bailey S, et al. Opioid Prescription Expenditures by Oregon Medicaid Enrollees after the 2014 Medicaid Expansion. Oral presentation presented at 46th Annual Meeting of the North American Primary Care Research Group; Nov, 2018; Chicago, IL. Marino M. Assessing the Impact of Health Policy Changes in Primary Care. Joint Statistical Meeting J, Vancouver, Canada. Assessing the Impact of Health Policy Changes in Primary Care. Oral presnetation presented at Joint Statistical Meeting; July 28-August 2, 2018; Vancouver, Canada. Marino M, Valenzuela S, Raynor L, et al. Validation of Medicaid Coverage in Electronic Health Records to Monitor Health Reform. WNAR Conference; June 27, 2017, 2017. Marino M, Valenzuela S, Hoopes M, DeVoe JE. A Two-Stage Logistic Regression Model to Identify Covariates that are Predictive of Agreement Between Insurance Sources. Oral Presentation presented at 10th International Chinese Statistical Association International Conference; Dec. 19-22, 2016; Shanghai, P. R. China.

23

Marino M, Killerby M, Hatch B, et al. Change in chronic disease biomarkers over time: The role of Medicaid Expansion. Oral Presentation presented at 44th annual meeting of the North American Primary Care Research Group; 2016; Colorado Springs, CO. Marino M, Killerby M, Angier H, et al. Using Electronic Health Records for Accurate Adult Medicaid Coverage Continuity. 44th annual meeting of the North American Primary Care Research Group; 2016; Colorado Springs, CO. Marino M, Huguet N, Springer R, Angier H, Hoopes M, DeVoe JE. Studying the Impact of the Affordable Care Act Using Electronic Health Records. Oral Presentation presented at 12th International Conference on Health Policy Statistics; Jan 10-12, 2018, 2018; Charleston, SC. Marino M, Angier H, Valenzuela S, et al. The Impact of the Affordable Care Act on Diabetes Biomarker Control. Oral Presentation presented at AcademyHealth Annual Research Meeting; June 24-26, 2018, 2018; Seattle, WA. Marino M, Angier H, Springer R, et al. The Affordable Care Act - Effects of Medicaid on Diabetes Biomarker Control. Oral Presentation on Completed Research presented at 45th annual meeting of the North American Primary Care Research Group; Nov 17-21, 2017; Montreal, Canada. Marino M, Angier H, Hoopes M, et al. Combining Historical Data and Propensity Score Methods in Observational Studies to Improve Internal Validity. Poster Presentation presented at American Statistical Association Conference on Statistical Practice; Feb 15-17, 2018, 2017; Portland, OR. Marino M. Using Electronic Health Records for Observational Cancer Studies. Invited Speaker presented at OHSU Knight Cancer Biostatistics Symposium; May, 2018, 2018; Portland, OR. Krollenbrock A, Baker D, Crosland K, et al. Developing a multi-level model for patient engagement in a Practice-based Research Network. 8th International Conference on Patient- and Family- Centered Care; June 2018, 2017; Baltimore, MD. Huguet N, Valenzuela S, Marino M, et al. Who remained uninsured after Medicaid expansion? Oral Presentation on Completed Research presented at 45th annual meeting of the North American Primary Care Research Group; Nov 17-21, 2017; Montreal, Canada. Hoopes M, Schmidt T, Winters K, et al. Prevalence and Characteristics of Cancer Survivors in Outpatient Safety Net Clinics. Poster Presentation presented at AcademyHealth Annual Research Meeting; June 24-26, 2018, 2018; Seattle, WA. Hoopes M, Angier H, Suchocki A, Dickerson K, DeVoe JE. Impacts of the Affordable Care Act Insurance Expansions on Community Health Centers: Current and Future Directions. OCHIN Learning Forum; April 14-16, 2016; Portland, OR. Hoopes M, Angier H, Gold R, et al. Impact of the First Year of Affordable Care Act Insurance Expansions on Community Health Center Encounters. 43rd annual meeting of the North American Primary Care Research Group; October 24-28, 2015; Cancun, Mexico.

24

Holderness H, O'Malley JP, Marino M, et al. Initial Source of Outpatient Services after Implementation of the Affordable Care Act Medicaid Expansion in Oregon. Poster presentation presented at OHSU Research Week; 2018, 2018; Portland, OR. Holderness H, O'Malley JP, Marino M, et al. Initial Source of Outpatient Services after Implementation of the Affordable Care Act Medicaid Expansion in Oregon. Poster Presentation presented at AcademyHealth Annual Research Meeting; 2018; Seattle, WA. Holderness H, O' Malley JP, Huguet N, et al. Site of First Primary Care Visit by Medicaid Enrollees after Implementation of the Affordable Care Act Medicaid Expansion. Oral Presentation presented at 46th Annual Meeting of the North American Primary Care Research Group; Nov 9-13, 2018, 2018; Chicago, IL. Hatch B, Tillotson C, Hoopes M, et al. Clinic factors associated with receipt of needed preventive services. 44th annual meeting of the North American Primary Care Research Group; 2016; Colorado Springs, CO. Ezekiel-Herrera D, Angier H, Springer R, et al. Modeling and Interpreting Three-Way Interactions for Disparities Research. Poster Presentation presented at American Statistical Association Conference on Statistical Practice; Feb 15-17, 2018, Rejected 2017; Portland, OR. DeVoe JE. Using Electronic Health Record Data from Community Health Centers to measure Utilization One Year Before and After Affordable Care Act Insurance Expansions in the United States. 3rd International Primary Health Care Reform Conference; March 17, 2016; Brisbane, Australia. DeVoe JE. Innovation and Discovery in Primary Care: Creating 21st Century Laboratories and Classrooms for Improving Health. Society for Teachers of Family Medicine Annual Spring Conference; May 2, 2016; Minneapolis, MN. DeVoe JE. Health Policy Roundtable: U.S. Healthcare System in Transition. Health Policy Roundtable, Australian Commonwealth Department of Health; March 22, 2016; Canberra, Australia. DeVoe JE. Healthcare Policy Seminar Series: U.S. Healthcare System in Transition: Before and After Obamacare. Healthcare Policy Seminar series, Australian Commonwealth Department of Health; March 22, 2016; Canberra, Australia. Angier H, Huguet N, Hoopes M, et al. Prevalence and Type of Pre-existing Conditions Before and After the Affordable Care Act among Vulnerable Populations. Oral Presentation on Completed Research presented at 46th Annual Meeting of the North American Primary Care Research Group; Nov 9-13, 2018, 2018; Chicago, IL. Angier H, Hoopes M, Marino M, Heintzman J, Huguet N, DeVoe JE. Impact of the Affordable Care Act on Racial and Ethnic Disparities in Medicaid Expansion vs. Non-Expansion States. 44th annual meeting of the North American Primary Care Research Group; 2016; Colorado Springs, CO.

25

Regulatory Requirements

Institutional Review Board (IRB) A separate Regulatory Binder containing all IRB and other regulatory documents is maintained by the OHSU Project Manager. Refer to this binder for more details about regulatory documents.

Approval See Appendix 1

Protocol See Appendix 2 Data Use Agreements and Amendments

Table 3. Data Use Agreements Parties Document Purpose Execution OCHIN and OHSU DUA Allows OCHIN to share limited data

set with OHSU 4-6-16

OCHIN and HCN DUA Allows OCHIN to share HCN data found in the ADVANCE CDM with OCHIN as a limited data set

6-3-16

OCHIN and OHSU DUA Amendment

Adds Community Vital Signs data to original DUA

5-22-17

OCHIN and Oregon DAA Allows OCHIN to use state-owned Medicaid data

6-20-16

See Appendix 3 for full agreements.

26

Decision Log

Table 4. Decision Log Study Influence Date of

Decision Decision Context

Decision

PACE Insurance Validation

Paper (Miguel)

03/29/16 PPD Meeting

Will exclude Medicare/Medicaid dual eligibles (since we cannot determine who they are in the ADVANCE data).

PACE Insurance Validation

Paper (Miguel)


Will exclude visits coded as being paid by another public payor – these don’t provide information about Medicaid coverage.

PACE Data Methods


Data Pull will include all states, however down the road if there are complications based on a state’s data they will be excluded.

PACE Bio Marker Paper

(Miguel)


Analysts will use Oregon Medicaid administrative data to create insurance cohorts to validate ADVANCE visit based insurance groups.

PPD Definitions


Patient state and clinic state will be given, classified as expansion vs non expansion. Patient state will be state at the patient’s most recent encounter, not historical state at date of encounter.

PPD Data Methods

03/25/16 Journal Club

Currently, the methods used within the "Propensity score weighting with multilevel data" article by Li, Zaslavsky and Landrum will be applied to all papers.

PPD Data Methods

04/01/16 Allison/ Jean/ Megan

Base data pull: Clinic inclusion criteria OCHIN: dept type=primary care, public health, or medical specialty specified as internal med, peds, or women's health. School-based health centers (SBHCs) included. Health Choice network: site type=CHC, peds, SBHC, or hospital specified as outpatient clinic. Exclude based on facility names including radiology, mental, behavioral, substance, dental, pharmacy, or lab. Facility type categorized into: primary care, public health, pediatrics, women's health, mobile, specialty, and SBHC.

27

PPD Data Methods

04/01/16 Allison/ Jean/ Megan

Base data pull: the following algorithm was used to assign go-live dates to HCN facilities: - PM golive = earliest encounter date per facility that is >=1/1/2003 (earliest actual go-live date for an HCN facility is in 2003) - Identify earliest vitals date that is >=1/1/2003 --> if earliest vitals date is >= PM golive date then EHR golive=earliest vitals date; else if earliest vitals date < PM golive then EHR golive=PM golive; else if no vitals date, then EHR golive = null.

PPDA Data Methods

04/25/16 Analyst Meeting

Blood glucose and HbA1c both in consideration for diagnosis of diabetes (included in Nichols algorithm).

PPDA Data Methods

05/02/16 Email ‘Encounters’ in the OCHIN system can include non- face-to-face contacts such as telephone visits, mychart updates, etc. In past studies, we have typically considered ‘visits’ to be the face-to-face and in-clinic subset of ‘Encounters’, though this definition has varied across studies/analyses. At times ‘encounter’ and ‘visit’ has been used interchangeably.

PPDA Data Methods

05/02/16 Email Other Ambulatory Visit (OA): Includes other non-overnight ambulatory encounters such as dental visits, hospice visits, home health visits, skilled nursing visits, other non-hospital visits, as well as telemedicine, telephone and email consultations. May also include "lab only" visits (when a lab is ordered outside of a patient visit), "pharmacy only" (e.g., when a patient has a refill ordered without a face-to-face visit), "imaging only", etc.

PPDA Data Methods

05/02/16 Email Ambulatory Visit (AV): Includes visits at outpatient clinics, physician offices, same day/ambulatory surgery centers, urgent care facilities, and other same-day ambulatory hospital encounters, but excludes emergency department encounters. Not limited to medical visits (e.g., includes mental/behavioral health visits).

PPDA Data Methods

05/02/16 Email If additional encounter details, such as provider type, facility type, level of service, etc., are used in the inclusion/exclusion criteria, they should be clearly specified in the outlines (i.e. “OB qualifying encounter types”: face-to-face with x diagnoses, y procedures, provider type(s), etc.).

28

PPDA Data Document

ation


High-level patient inclusion/exclusion criteria for each paper will be tracked in columns in Manuscript Tracker; more detailed documentation (e.g., 'readme' and data dictionaries) will be provided with each study dataset and included as an attachment in the manuscript tracker.

PPDA Data inclusion


Residents of states that are not included in study (but were seen at eligible facilities in one of our 19 states) will be excluded.

PPDA Data Methods


There may be data errors in the provider_type variable in the encounters file (i.e., an AV visit may have dentist listed as provider type – which one is correct?) – safer to use AV count based on ADVANCE data team’s large amount of work on this categorization.

PPDA Data Definition

Email PACE: Pre=2013, Post=2017/PREVENTD: Pre=2012,

Post=2020 PPDA Data

inclusion 06/16/16 Manuscript

Meeting Keep continually insured, uninsured in pre and mix of all insured visits, track how many visits were paid by Medicaid (will be able to isolate based on persons insurance status), throw out public insurance/Medicare patients. Patient with continual private insurance in periods will be continuously insured.

PPDA Data inclusion

06/16/16 Manuscript Meeting

Biomarker change could change based on preexisting condition in Medicare group, Patients with Medicare record will be excluded (all papers).

PPDA Data analysis

8/18/16 Manuscript Meeting

Include WI as an expansion state in all analyses.

PPDA Data Methods

06/14/17 Manuscript meeting

HCN data for "smoking" not specific enough; using tobacco-only analysis instead of separating out type of Tobacco use (unless able to code further).

PPDA Data Methods

06/14/17 Manuscript meeting

Collect pre-AND post- Charlson Comorbidity Index score (using problem list)

PPDA Data analysis

1/4/18 PPDA Manuscript Meeting

ADVANCE data from 2017 to be pulled February 2018.

PPDA Data analysis

2/14/18 Other Refer to as “Enhanced Charlson Comorbidity Index” and cite: Charlson ME, Charlson RE, Marinopoulos SS, Briggs WM, Hollenberg JP. The Charlson Comorbidity Index is adapted to predict costs of chronic disease in primary care patients. J Clin Epidemiol. 2008; 61:1234-40.

29

Appendix 1. IRB Approval

APPROVAL OF SUBMISSION

May 1, 2017 Dear Investigator:

On 5/1/2017, the IRB reviewed the following submission:

IRB ID: IRB00011858 MOD or CR ID: MOD00007213 Type of Review: Expedited-Minor Modification

Title of Study: Impacts of the Affordable Care Act Title of modification Protocol Update

Principal Investigator: Jennifer Devoe Funding: Name: DHHS CDCP, PPQ #: 1007159, Funding Source:

U18 DP006116; Name: DHHS NIH Natl Cancer Inst, PPQ #: 1007752, Funding Source: pending; Name: DHHS Agency for Hlth Care Policy & Rsch, PPQ #: 1006693, Funding Source: R01HS024270

IND, IDE, or HDE: None

The IRB granted final approval on 5/1/2017. The study is approved until 3/30/2018.

30

Review Category: Expedited-Minor Modification

Copies of all approved documents are available in the study's Final Documents (far right column under the documents tab) list in the eIRB. Any additional documents that require an IRB signature (e.g. IIAs and IAAs) will be posted when signed. If this applies to your study, you will receive a notification when these additional signed documents are available.

Ongoing IRB submission requirements:

• Six to ten weeks before the expiration date, you are to submit a continuing review to request continuing approval.

• Any changes to the project must be submitted for IRB approval prior to implementation. • Reportable New Information must be submitted per OHSU policy. • You must submit a continuing review to close the study when your research is

completed.

Guidelines for Study Conduct

In conducting this study, you are required to follow the guidelines in the document entitled, "Roles and Responsibilities in the Conduct of Research and Administration of Sponsored Projects," as well as all other applicable OHSU IRB Policies and Procedures.

Requirements under HIPAA

If your study involves the collection, use, or disclosure of Protected Health Information (PHI), you must comply with all applicable requirements under HIPAA. See the HIPAA and Research website and the Information Privacy and Security website for more information.

IRB Compliance

The OHSU IRB (FWA00000161; IRB00000471) complies with 45 CFR Part 46, 21 CFR Parts 50 and 56, and other federal and Oregon laws and regulations, as applicable, as well as ICH-GCP codes 3.1-3.4, which outline Responsibilities, Composition, Functions, and Operations, Procedures, and Records of the IRB.

Sincerely,

The OHSU IRB Office

31

Appendix 2. Project Protocol

Protocol Title Impacts of the Affordable Care Act

Objectives

The purpose of this study is to understand the impacts of the Affordable Care Act (ACA) in a safety net population. We will assess the following in states that expanded Medicaid compared to those that did not or among states that expanded Medicaid:

• Pre-post rates of health insurance status and payer mix • Pre-post access to healthcare (i.e., visits, provider workforce, receipt of recommended

healthcare services, receipt of chronic disease specific care and outcomes; receipt of preventive care, appropriate care for specialized populations such as cancer survivors and patients with multimorbidity/complex healthcare needs)

• Pre-post healthcare expenditures • Pre-post changes in health disparities (e.g., low-income, racial and ethnic minorities,

rural versus urban) We will look at differences in access to healthcare, expenditures, and disparities among those who gained coverage in the post-period, those who already had coverage, those with intermittent coverage, and those who remained uninsured. We will also try to understand how aggregated measures of social determinants of health constructed from community-level geocoded data (i.e., fast food availability, parks, income distribution, and racial/ethnic breakdown) from publicly available sources, such as the US Census Bureau, change the impact of disparities, access, and receipt of healthcare in relation to gaining or not gaining health insurance. We propose to investigate how complex, and early-onset diseases affect the receipt of healthcare services later in life. Some of these analyses will focus on sub-group populations including patients with a history of cancer and those with multiple or complex chronic conditions. To accomplish these analyses, we will examine data from patients of all ages as many childhood cancers have a high likelihood of survival, and it is important to identify these patients as they are likely to have additional healthcare needs throughout their lifespan.

Background Health insurance facilitates access to care and reduces unmet healthcare needs,1-4 yet 47 million Americans had no coverage in 2012.5 The ACA, the largest healthcare-related legislation in the United States (US) since Medicare’s establishment in 1966, was enacted with the goal of expanding coverage to all citizens and legal residents.6 The ACA increases opportunities to gain health insurance, including expansions in Medicaid coverage to individuals earning ≤ 138% of the federal poverty level (FPL) and the establishment of health insurance marketplaces. In 2012, the Supreme Court ruled that states were not legally required to implement Medicaid

32

expansions, and those opting out could not be penalized.7 As of December 2014, 27 states (and D.C.) implemented the expansion;8 the number of states is increasing. Many persons directly affected by these expansions are seen at community health centers (CHCs), which comprise much of our nation’s healthcare ‘safety net’ and serve a rapidly increasing number of patients;9 in 2012, 36% of CHC patients were uninsured.10 The effect of the ACA on health insurance coverage and access to health care services is unknown, thus it is imperative to assess its impact.

Study Design Secondary data analysis. We will use electronic health record (EHR) data from the Accelerating Data Value Across a National Community Health Center Network (ADVANCE) clinical data research network (CDRN) data warehouse and Oregon Medicaid administrative data. The ADVANCE CDRN is a unique ‘community laboratory’ for research with underrepresented populations that includes patients receiving care in safety net clinics. The ADVANCE CDRN data warehouse includes integrated longitudinal outpatient EHR data from several organizations, including OCHIN, Health Choice Network, Fenway Health, and Legacy Health. Oregon Medicaid administrative data will be linked to OCHIN Oregon clinics in the ADVANCE CDRN to assess the receipt of healthcare services outside the ADVANCE network clinics and all healthcare expenditures. Additionally, we will partner with the state cancer registries of one or more states to complete probabilistic linkage between our patient population and reported cancer cases. It is currently unknown how well personal cancer history and related data are documented in the outpatient EHR (i.e., data contained in the ADVANCE CDRN), thus our ability to identify and characterize cancer survivors may be incomplete. A linkage with one or more state cancer registries will supplement the CDRN data by comprehensively identifying cancer survivors in our cohort. This will allow us to test and validate our EHR-based algorithms, and to evaluate post-cancer care services provided by community health centers and the impact of insurance coverage for this population.

Study Population a) Number of Subjects Approximately 2.3 million total possible subjects. All analyses will contain a subset of these

subjects.

b) Vulnerable Populations Children ages 0-18 will be included for methods development and analyses related to childhood cancer and long-term care for cancer survivors. There is little literature about childhood cancers in vulnerable populations (e.g., economically marginalized, uninsured, or Medicaid insured) and associated disparities. Furthermore, adults with a history of childhood or adolescent cancer may require special follow-up care and screening

33

throughout their adult lives, thus we wish to have access to their EHR records from earlier ages where available.

c) Setting This research will be conducted in collaboration with OCHIN. OCHIN requests the OHSU IRB serve as the IRB of record for this study.

d) Recruitment Methods and Consent Process All patients with data in the ADVANCE CDRN may be included in this study. Subsets may include only OCHIN EHR data, only Oregon OCHIN EHR data, only persons with certain diseases or conditions (e.g., cancer, diabetes, obesity), etc. For analysis of expenditures, Oregon OCHIN clinics from the ADVANCE CDRN will be linked to Oregon Medicaid administrative data using Medicaid identification number; subjects found in both datasets will be included. Similarly, for cancer registry linkage analyses only patients identified in both datasets will be included in analyses. The ADVANCE data was collected under a waiver of authorization as the data warehouse poses very little risk to patients, and it was not practical to consent the number of patients included in the dataset.

Procedures

Secondary data analysis will be conducted.

Data and Specimens a) Sharing of Results with Subjects Results will not be directly shared with subjects.

b) Data and Specimen Banking No specimens are being collected. Data is not being banked for future research.

Data Analysis

We will summarize baseline measures using descriptive statistics and data visualization methods (e.g., histograms, scatter plots) to characterize baseline data across clinic and state groups. We will estimate the differences in access and receipt of recommended health care. We will also assess health care expenditures. Our primary methodological approach will utilize difference-in-differences (DID) methodology. The DID approach has been frequently used by health economists and health services researchers to account for potential secular effects and changing policies that would affect both expansion and non-expansion states over time, while adjusting for potential confounders.11-16 We will implement state random effects in clinic-level analyses and both clinic and state random effects in patient-level analyses to control for correlation of observations nested in clusters (e.g., individuals nested in clinics which are nested in states). We recognize that the assumption for random effects may not be met and we will also assess the robustness of our

34

assumptions by running models that treat clinics and states as fixed effects. We will formally test the validity of the random effects using the standard Hausman test,17 as well as assessing overall qualitative differences that may arise between the random and fixed effects models. We will use GLMM models,18 which offer flexible regression modeling to accommodate different sources of correlations (serial, intra-clinic, and intra-state), categorical and continuous covariates, and fixed and time-dependent covariates. These methods offer a wide range of parametric distributions to model the dependent variables, including logistic regression (binary data), beta regression (percent data), Poisson regression (count data), and Gaussian regression (normally distributed data). The distribution of the outcomes of interest will be examined before selecting an analysis model; specific models will be refined through an iterative process. Potential variables included in the analyses Description

Age Categories Sex Male/female, other

Race White, American Indian/Alaskan Native, Asian, Black, Native Hawaiian/Pacific Islander, Other

Ethnicity Hispanic/non-Hispanic Household Income Federal poverty level; overall income; % FPL Language Preference English, Spanish, other, etc. Clinic Information Rural/urban; panel size; panel demographics; provider characteristics Medicaid Expansion Status ACA Medicaid expansion state/non-expansion state; date of expansion

Visits Frequency and type of visits Healthcare Need Special need because of a chronic condition

Insurance Coverage status and type at visit; changes in status and type over time; % covered over time

Continuity of Care % of visits at the same site and with same provider; Continuity of Care Index

Service Utilization Number and types of all billed encounters overall and yearly; all services received; disease specific services received; preventive services received; amount of time services are delayed

Expenditures Categories; overall

Health outcomes Biomarkers such as HbA1c, blood pressure, body mass index, cholesterol measurements, etc

Diagnoses Number and type of chronic and acute conditions

Community Vital Signs Community markers of social determinants of health (e.g., fast food availability, parks, income distribution, racial/ethnic breakdown, healthy food availability, income distribution, homeownership)

Cancer registry elements

Date of diagnosis, primary type, stage, sequence, histology, behavior, grade, insurance at diagnosis, family history, treatment, vital status, date of death

35

Privacy, Confidentiality and Data Security

We will obtain EHR data for patients included in the ADVANCE CDRN data warehouse. Data has already been transferred to the ADVANCE data warehouse from OCHIN, HCN, Fenway Health, and Legacy Health. The ADVANCE data are stored securely at OCHIN. All of the protections currently in place for the ADVANCE data warehouse will apply to this study. We will request the Oregon Medicaid administrative data from the Oregon Health Authority and collect it under a Data Authorization Agreement. The data transferred from OCHIN to OHSU will be de-identified and shared under a Data Use Agreement. The dataset will be de-identified by assigning each patient a unique subject identification that is not based on any personal identifiers. All data will be stored on secure servers and password protected computers. All data will be computerized and managed on HIPAA-compliant computers. All data work will be done on password-protected, HIPAA-compliant computers. All data will be stored and backed up on password-protected secure servers. All reports will describe results in aggregate form only. For cancer registry linkage activities, we will access personal identifiers sufficient for completing probabilistic matching; direct identifiers will be stripped from all datasets and replaced with anonymized patient identifiers at the completion of linkage and for the duration of analyses. Data access and linkage procedures will be governed by Data Access Agreements between each relevant state agency and OCHIN. Any data transferred from OCHIN to OHSU will fall under the same protocol as outline in the Data Use Agreement between these two institutions.

Risks and Benefits a) Risks to Subjects

There is a small risk of a loss of confidentiality.

b) Potential Benefits to Subjects There is no direct benefit to the subject for taking part in this study.

References

1. Asplin BR, Rhodes KV, Levy H, et al. Insurance status and access to urgent ambulatory care follow-up appointments. JAMA. 2005;294(10):1248-1254.

2. Smolderen KG, Spertus JA, Nallamothu BK, et al. Health care insurance, financial concerns in accessing care, and delays to hospital presentation in acute myocardial infarction. JAMA. 2010;303(14):1392-1400.

3. Burstin HR, Lipsitz SR, Brennan TA. Socioeconomic status and risk for substandard medical care. JAMA. 1992;268(17):2383-2387.

4. Bindman AB, Grumbach K, Osmond D, et al. Preventable hospitalizations and access to health care. JAMA. 1995;274(4):305-311.

5. Kaiser Commission on Medicaid and the Uninsured. Key Facts about Health Insurance on the Eve of Health Reform. Menlo Park, CA.2013.

36

6. Henry J. Kaiser Family Foundation. Summary of the Affordable Care Act. Menlo Park, CA2013.

7. Supreme Court of the United States. National Federation of Independent Business v Sebelius. 2012; http://www.supremecourt.gov/opinions/11pdf/11-393c3a2.pdf. Accessed January 15, 2015.

8. The Henry J Kaiser Familty Foundation. Status of state action on the Medicaid expansion decision. 2014; http://kff.org/health-reform/state-indicator/state-activity-around-expanding-medicaid-under-the-affordable-care-act/. Accessed January 7, 2015.

9. Morgan D. Health centers for poor, uninsured see ranks swell. 2012; http://www.reuters.com/article/2012/05/01/us-usa-healthcare-centers-idUSBRE8401JL20120501?feedType=RSS&feedName=everything&virtualBrandChannel=11563. Accessed January 15, 2015.

10. National Association of Community Health Centers. A Sketch Of Community Health Centers, Chart Book 2014. Bethesda, MD. 2014.

11. Bertrand M, Duflo E, Mullainathan S. How Much Should We Trust Differences-in-Differences Estimates? The Quarterly Journal of Economics. 2004;119(1):249-275.

12. Friedberg MW, Schneider EC, Rosenthal MB, Volpp KG, Werner RM. Association between participation in a multipayer medical home intervention and changes in quality, utilization, and costs of care. JAMA. 2014;311(8):815-825.

13. Higgins S, Chawla R, Colombo C, Snyder R, Nigam S. Medical homes and cost and utilization among high-risk patients. The American journal of managed care. 2014;20(3):e61-71.

14. Werner RM, Duggan M, Duey K, Zhu J, Stuart EA. The patient-centered medical home: an evaluation of a single private payer demonstration in New Jersey. Medical care. 2013;51(6):487-493.

15. Werner RM, Konetzka RT, Polsky D. The effect of pay-for-performance in nursing homes: evidence from state Medicaid programs. Health services research. 2013;48(4):1393-1414.

16. Donald S, Lang K. Inference with Difference-in-Differences and Other Panel Data. The Review of Economics and Statistics. 2007;89(2):221-233.

17. Hausman JA. Specification tests in econometrics. Econometrica. 1978;46:1251-1271. 18. Dean CB, Nielsen JD. Generalized linear mixed models: a review and some extensions.

Lifetime Data Anal. 2007;13(4):497-512.

37

Appendix 3. Data Use Agreements

A Study Manual for: The PACE Study · from 2014 to 2015 who were either newly, returning, or...

Documents

Transcript of A Study Manual for: The PACE Study · from 2014 to 2015 who were either newly, returning, or...